Clinical decision support system for breathlessness

This review summarizes high-level evidence on clinical decision support systems, both more classical rules-based and emerging artificial intelligence-based examples. It discusses their potential and concerns, how to evaluate them and describes future directions based on published evidence, clinical experience and broader experience working with others in this space.

There is significant potential for CDSS to benefit elicitation and assessment of breathlessness, assist clinical decision making, improve interpretation of common diagnostic tests for breathlessness and personalize management and patient education. Even so, current evidence from trials and real-world studies in this space remains limited with multiple studies ongoing. However, evidence from studies of model development indicates that CDSS have clinically acceptable performance levels for differentiating breathlessness causative conditions such as COPD, heart failure, lower respiratory tract infection or combinations of these. The extent to which this evidence translates to real world clinical benefits remains unknown.

In the era of CDSS leveraging significant volumes of data, there is the potential to augment the less precise nature of clinician prediction with that of AI prediction. This, combined with clinical judgment can support better care for patients and populations. CDSS are likely to be particularly valuable in settings with workforce constraints though we must also remember that there are limitations to CDSS use and applicability. It remains important to support and undertake high quality studies testing these tools in clinical practice.